The Effect of Seasonal Adjustment on Turning Points Detection

• Activity Number: 516
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Gaetana Montana*+ and Gian Luigi Mazzi and Monica Billio
• Companies: EUROSTAT and EUROSTAT and Università di Venezia
• Keywords: Seasonal Adjustment ; non linear models ; Coincident indicators ; Comparative analysis ; turning points detection

Abstract

Monitoring the cyclical economic situation requires a timely and accurate detection of turning points. Eurostat produces two coincident indicators for turning points detection of business and growth cycle based on univariate Markov Switching model fitted to each component of the two indicators. The value of the indicator is a weighted average of the filtered probabilities of each component. The indicators are based on Eurostat seasonally adjusted data. This paper investigates whether alternative seasonal adjustment methods affect the detection of turning points. We have then compared the results of the two indicators obtained when components are seasonally adjusted by seven different methods. The application shows that while most common seasonal adjustment methods do not affect significantly recession signals, moving towards non standard methods shifts the location of turning points.